1 INTRODUCTION
3.5.3 Study IV
The BBS showed a ceiling effect and was unable to predict falls in Study II. The Mini-BESTest appeared to be an alternative.50,59,74 In order to make a comparison of the two balance scales, 46 individuals able to walk 10 metres completed both the BSS and Mini-BESTest.
The floor and ceiling effects of the BBS and Mini-BESTest were calculated and confirmed if more than 15 per cent of the individuals scored either the lowest or highest possible total score.66
The internal consistency of the BBS and Mini-BESTest was measured using Cronbach‟s alpha, where an alpha value between 0.70 and 0.80 was regarded as satisfactory. To investigate the construct validity, a Mann–Whitney U-test was used to compare the subgroups for differences, and Spearman‟s rank correlation coefficient (rs) was used to check for correlations between continuous variables that were not normally distributed. Correlation coefficients were interpreted as follows: low 0.26‒0.49, moderate 0.50‒0.69, high ≥0.70.61
Receiver operating characteristics (ROC) were used to determine the relative ability of the BBS and Mini-BESTest to discriminate known groups: high versus low functional score dichotomised at median, community walkers not using aids versus all other participants, low frequency versus recurrent falls (0‒2/>2) and low versus high concerns about falling dichotomised at median.
Further, Rasch analysis was used to evaluate the internal validity of the rating scales of the BBS and the Mini-BESTest. The category structure of the balance scales were examined with values for observed average, the order of categories, outfit mean square (MnSq) values below 2.0, and by observing the probability curves.
3.5.3.1 Additional analysis
Bland‒Altman method was used to investigate the limits of agreement between BBS and Mini-BESTest.
31
4 RESULTS PARTICIPANTS 4.1
Characteristics of the patients in the different studies are shown in Table 5.
Of the 224 individuals included in Study I, 151 were defined as wheelchair users and 73 as ambulatory. The cut-off was determined, with inspiration from Boswell-Ruys et al.,69 according to the participants‟ own ratings of the time ratio between wheelchair use and walking to meet mobility needs (wheelchair/ambulating: 0/100, 75/25, 50/50, 25/75 or 100/0). Five participants who stated an equal ratio were discussed and finally classified as ambulatory by the SCIP FALLS Study research group.The ambulatory ratio in Norway was 42 per cent and in Sweden 25 per cent. The total sample consisted of individuals with great variations in terms of their extent of injury, from individuals with little remaining sensorimotor function who were reliant on powered wheelchairs to individuals with almost normal function and no need of aids.
Of the 73 included ambulatory participants, 68 completed the 12-month follow-up for falls, a dropout rate of seven per cent. The reasons for dropouts were three deaths, one withdrawal and one case of insufficient fall data.
Study III was carried out on a Norwegians sample. Eleven participants from Study II were invited to participate, considering that their places of residence were within a reasonable distance to allow face-to-face interviews, and ensuring a variety in age, injury and functional levels and socio-demographic variables. One elderly individual (> 80 years of age) declined to participate due to poor health. To ensure information power102 and to add to the variety of experiences in regard to falls, five additional individuals were recruited after the recruitment for Study II was completed. Altogether, the sample in Study III was comprised of 15 informants.
All 47 participants from the Norwegian cohort of the SCIP FALLS Study in Study I, who were able to walk 10 metres independently with or without walking aids, were included in the sample for Study IV. One participant failed to perform the Mini-BESTest and was thus excluded.
32
Table 5. Characteristics of the participants in all the studies. Values are presented as n (%), or as otherwise stated.
Characteristics Study I
(n=224)
Study II (n=68)
Study III (n=15)
Study IV (n=46)
Male 173 (77) 45 (66) (67) 32 (70)
Age, mean years (SD) 50 (15) 55 (15) 47(18) 55 (17)
Educational level
Secondary school or less High school
College/university
73 (33) 66 (30) 83 (37)
23 (34) 20 (29) 25 (37)
5 (33) 4 (27) 6 (40)
21 (46) 11 (24) 14 (30) Work
Not working Working
123 (55) 101 (45)
41 (60) 27 (40)
10 (67) 5 (33)
37 (80) 9 (20) SCI characteristics
Duration of injury
Median years (min‒max) 15 (1‒56) 12 (1‒40) 7 (2‒34) 7 (1‒41) Injury level
Cervical
Thoracic (T1‒6) Thoracic (T7‒12) Lumbar
114 (51) 33 (15) 59 (26) 18 (8)
42 (62) 1 (2) 17 (25) 8 (12)
7(47) 0 (0) 3(20) 4 (27)
30 (65) 1 (2) 7 (15) 8 (17) Completeness of injury1
AIS A AIS B AIS C AIS D
100 (45) 31 (14) 20 (9) 73 (32)
2 (3) 1 (2) 3 (4) 62 (91)
0 (0) 0 (0) 2 (13) 13(87)
1(2) 1 (2) 3 (7) 41 (89)
LEMS2, median (min‒max) 0 (0‒50) 45 (19‒50) ‒ 46 (20‒50)
Injury mechanism Sport
Violence Traffic Fall Other
51 (23) 3 (1) 91(41) 69 (32) 10 (4)
14 (21) 0 (0) 22(33) 28 (39) 4 (6)
3 (20) 0 (0) 8 (53) 3(20) 1(7)
9 (20) 1 (2) 13 (28) 21(46) 2 (4) Functional characteristics
SCIM motor3, median (min‒max) 19 (3‒40) 39 (19‒40) ‒ 37 (8‒40)
WISCI4, median (min‒max) ‒ 20 (12‒20) ‒ 20 (9‒20)
1AIS =American Spinal Injury Association impairment scale. A= complete injury, B= sensory incomplete injury, C= Motor incomplete injury; more than half of the key muscles have muscle grade <3, D= motor incomplete injury;
at least half of the key muscles have muscle grade ≥3.
2LEMS=Lower extremity motor score according to International Standard Neurological Classification of Spinal Cord Injury, sum score 0‒50 points.
3SCIM = Spinal Cord Independence Measure III, motor items sum score 0‒40 points.
4Walking Index for Spinal Cord Injury II, sum score 0‒20 points.
33
FALLS AND ASSOCIATED FACTORS IN INDIVIDUALS WITH SCI 4.2
Seventy-six per cent of the total sample (n=224) reported falling and 51 per cent reported recurrent (>2) falls in the previous year. The final multiple regression model showed that ambulatory individuals were 2.9 times more likely to report recurrent falls compared to wheelchair users.
Further, those who were able to get up from the ground by themselves were 2.2 times more likely to report recurrent falls than those who were not able to do so. In addition, those who exercised regularly for at least 30 minutes once or twice a week were 1.9 times more likely to report recurrent falls compared to those who exercised less. With increasing age, the odds of reporting recurrent falls decreased 3 per cent per year (Table 6).
As mode of mobility was the factor associated with the highest OR for recurrent falls and the distinction is clinically important, separate subgroup models were run for wheelchair users and ambulatory individuals. For ambulatory individuals, the odds of reporting recurrent falls were 4.7 times greater than for those who were able to get up by themselves from the ground. Further, the odds decreased 24 per cent per increasing scale-step on the International SCI Quality of Life Scale (i.e. with increasing quality of life).
Table 6. The final multiple logistic regression models for the total sample and for the ambulatory subgroup showing factors associated with recurrent (>2) falls. First category is a reference for the categorical variables unless otherwise is stated. No decimals indicated if OR≥10
Variable β OR 95% CI p–value
Total sample
Age 0.12 0.97 0.95–0.99 0.012
Gender (ref. woman) 0.48 1.61 0.80–3.24 0.182
Wheelchair or ambulatory 1.08 2.93 1.43–6.03 0.006
Able to get up by one–self (no/yes) 0.77 2.15 1.16–3.99 0.015 Exercise previous year (no/yes)* 0.64 1.90 1.09–3.38 0.029 Ambulatory subgroup
Age –0.00 1.00 0.96–1.03 0.850
Gender (ref. woman) –0.10 1.10 0.34–3.60 0.869
Able to get up by one–self (no/yes) 1.57 4.74 1.06–21 0.042
General quality of life** –0.28 0.76 0.59–0.99 0.045
Overall model fits (Hosmer and Lemeshow test):
Total sample: Chi2=6.135, df=8, n= 223, p= 0.632 Ambulatory subgroup: Chi2=6.953, df=8, n= 73, p=0.542
* No „regular‟ or „regular‟ exercise previous year. ** International SCI Quality of Life Basic Dataset.
34
FALLS, INJURIOUS FALLS AND THEIR PREDICTORS IN 4.3
AMBULATORY INDIVIDUALS WITH SCI
During the one-year registration period, 272 falls were reported, with sports-related falls excluded.
Twelve participants (18%) did not fall, 17 (25%) fell once, six (9%) fell twice and 33 (48%) fell three times or more. The median number of falls was 2 (Q1–Q3: 1–6), with the highest number being 23. The peak prevalence for falls was in the middle age range group (46–60 years).
At least one injurious fall was reported by 44 of the participants (65%), median number being 1 (Q1–Q3: 0–2). Women reported significantly more falls than men. Injurious falls were positively correlated with the number of reported falls (rs=0.74, p≤0.001). Of the reported falls 41% were injurious. Of the injurious falls 84% resulted in minor injuries, 16 per cent in moderate or serious injuries. Nine participants (13%) sustained a moderate injury. Three participants (4%), all women above the age of 55, sustained serious injuries.
Forty-six per cent of the falls occurred indoors. Most falls (60%) happened in the daytime, were spread throughout the week and occurred during a variety of activities. The activity that accounted for most indoor falls was getting in and out of a chair (20%). Walking on uneven or slippery surfaces accounted for most outdoor falls (31% and 21%, respectively). Although not reaching significance (p=0.054), more moderate and serious injuries tended to occur indoors (13 vs. 6) while more falls with minor injuries occurred outdoors (135 vs. 113).
Two data punching errors regarding the independent variable fear of falling were detected after Study II was submitted to the journal. However, when correcting for these errors, the final multiple logistic models for recurrent and injurious falls were not altered, but the numbers are somewhat changed. The corrected final multiple regression models for recurrent falls and injurious falls are presented below and in Table 7.
The final multiple logistic regression model (Table 7) showed that the odds of recurrent falls were at least 8.6 times higher (lower CI) for individuals with a history of recurrent falls in the previous year compared to those without such a history, and 6.1 times higher for individuals who were afraid of falling compared to those who had no such fear. With increasing time spent walking 10 metres, the odds of recurrent falls increased by 31 per cent per second. The wide CI and difficulties with estimating the true OR for recurrent falls in the previous year (as displayed in Table 7) were caused by the small sample size. The final model explained 47‒62 per cent of the variation in the dependent variable recurrent falls, and 87 per cent of all cases were correctly
35
Table 7. The final multiple logistic regression models showing predictors for recurrent (>2) falls and injurious falls in ambulatory individuals. First category is reference for categorical variables.
No decimals indicated if OR≥10.
Variable β OR 95% CI p-value
Recurrents falls
Ten Metre Walk Test, max. speed (s) 0.27 1.31 1.02–1.68 0.034 Falls previous year (0–2/>2) 4.71 111 8.58–1425 ≤0.001
Fear of falling (no/yes) 1.81 6.09 1.43–26 0.015
Injurious falls
Age (years) 0.04 1.04 0.99–1.08 0.093
Gender (man/woman) 1.00 2.71 0.68–11 0.158
Falls previous year (0–2/>2) 1.43 4.16 1.21–14 0.023
Fear of falling (no/yes) 1.45 4.25 1.26–14 0.018
Overall model fit (Hosmer and Lemeshow test):
Recurrent falls; Chi2=11.28, df=8, n=67, p=0.127. Cox &Snell R2= 0.47, Nagelkerke R2=0.62 Injurious falls; Chi2=6.58, df=8, n=68, p=0.133. Cox &Snell R2= 0.25, Nagelkerke R2=0.34
classified as recurrent fallers.The odds of injurious falls were 4.2 times higher for individuals with history of recurrent falls previous year compared to those without, and 4.3 times for individuals afraid of falling compared to those not afraid of falling (Table 7). The final model explained 25–
34 percent of the variation in the dependent variable; 77 percent of all cases were correctly classified as having an injurious fall.
4.3.1 Additional results
4.3.1.1 Time to first fall
When investigating falls in all 73 included participants, 57 (84%) fell at least once. By six months, 25 of the 71 (35%) participants that were still being followed had fallen three times.
The mean time to first fall was 19.7 (95% CI 15.3‒24.2) weeks, and to the third fall it was 37.5 (95% CI 33.2‒41.8) weeks (Figure 7).
4.3.1.2 Perceived reasons for falls
The participants perceived loss of balance, muscle weakness and hazards in the indoor and outdoor environments as main causes in 64 per cent of the fall incidents: hazards in 26 per cent of the outdoor and 10 per cent of the indoor falls and muscle weakness in 26 per cent of the indoor falls and 11 per cent of the outdoor falls. However, in 45 per cent of the falls, a combination of causes was reported. There was no significant correlation between balance control measured with
36
Figure 7. Kaplan–Meier plot showing time to first fall for all ambulatory individuals included in Study II. The dotted vertical line indicates 6 months (26 weeks) and the dotted horizontal line the 50 per cent probability of not falling (n=73).
BBS and recurrent falls (p=0.076), although there was a negative but low correlation between the number of falls and BBS (rs=-0.27, p=0.024).
4.3.1.3 Fall-related psychological aspects
At baseline, the mean score on concern about falling (FES-I) for the ambulatory group (n=73) was 29.4 (SD 8.2). There were no significant differences between participants younger than 65 years of age and those 65 years or older in the FES-I scores (p=0.591). Forty-six participants (63%) reported that they were afraid of falling, 19 (26%) were somewhat concerned, 18 (25%) were fairly concerned and 9 (12%) were very concerned. There were no significant differences between participants younger than 65 years of age and those 65 years or older in regard to reported fear of falling (p=0.512).
At the 12-month follow-up, 58 of the 73 included participants (79%) answered the FES-I questionnaire and 64 (88%) answered the question on fear of falling (yes/no). The mean score on the FES-I was 30.6 (SD 9.2). The mean paired difference was 1.5 points (95% CI -2.9 to -0.1), and the standard error of the mean (0.715) demonstrated a significant (p=0.040) difference. There was
37
significantly less reported fear of falling at follow-up compared to at baseline (49% vs. 56%, p=0.010).
4.3.1.4 Falls from a retrospective and a prospective perspective
There were significantly fewer falls registered falls (p<0.001) and significantly fewer recurrent fallers (p=<0.001) identified at the 12-month follow-up compared to the retrospectively reported falls and recurrent falls previous year at baseline (Figure 8). Table 8 displays the relationship between retrospectively reported and prospectively registered falls.
Figure 8. The 12-month fall incidence reported retrospectively and prospectively by the ambulatory individuals in Study II (n=68).
Table 8. Prospectively registered falls from baseline plotted against falls the previous year reported at baseline for the ambulatory individuals in Study II (n=68).
Falls previous year
0 1‒2 3‒5 >5
Prospectively registered
falls
0 3 3 2 4
1‒2 10 5 5 3
3‒5 0 1 10 2
>5 0 0 7 13
0 5 10 15 20 25 30 35 40
0 falls 1-2 falls 3-5 falls > 5 falls
Percentage of individuals
Retrospectively 12-month incidence
Prospectively 12-month incidence
100
38
INDIVIDUAL PERSPECTIVES ON FALLS AND FALL RISK 4.4
The analysis of the interviews with 15 ambulatory individuals revealed one overarching theme,
„Falls challenge identity and self-image as normal‟. This was comprised of two main themes interpreting the informants‟ experience of how falls, risk of falling and fall-related consequences were influenced by their incomplete SCI; „Walking with incomplete SCI involves minimising fall risk and fall-related concerns without compromising identity as normal‟ and „Walking with incomplete SCI implies a willingness to increase fall risk in order to maintain identity as normal‟.
The two main themes were formed by seven subthemes which represented the informants‟
contradictory view of falls; on one hand, the enhanced risk of falling and falls was considered to be a part of life, and the informants developed preventive strategies and justified their concerns to protect their vulnerable bodies against falls and fall-related consequences. On the other hand, they described a willingness to take risks to emphasise a normal identity (Figure 9).
Maintaining an identity and self-image as normal was important for the informants, and falling in everyday situations was perceived as embarrassing and threatened their self-image.
... it’s embarrassing... They all stand there watching, and you look like an idiot, and a whole lot of them think I’m drunk, you know. (woman, 43)
…you feel simply a bit handicapped, and you don’t want to admit that. (man, 39)
To avoid falls, they used several preventive measures, and many talked about their enhanced fall risk awareness and how they coped with challenging fall risk situations.
…the head becomes a sort of control freak in relation to the body. Yeah, so there’s a lot of head traffic, just like to go out for a walk (in the winter). (woman, 32)
Although trying to avoid falls, most informants regarded falls as a part of life and they tried to protect their vulnerable bodies by using protective strategies when falling to prevent injury.
... I try… to actually not put my hands in front of me, I try to protect my body, use my hands to protect my head, protect my belly, protect myself instead... (man, 29)
However, for a few informants, mostly women who had been injured in falls, falling was a big concern and restricted their activities. One 26-year-old woman having difficulties getting up from the ground independently stated, … I don’t go to places where there are no people. Losing control
Figure 9. Overview of results showing that perceptions of falls and fall risk in ambulatory individuals with incomplete SCI could be expressed in one overarching theme, two main themes, and seven subthemes.
Falls challenges identity and self-image as normal
Walking with incomplete spinal cord injury involves minimising fall risk and fall-related concerns without compromising
an identity as normal
Fall risk awareness
Fall-related psychological
issues
Preventive and protective
strategies
Dilemmas with aid use
Walking with incomplete spinal cord injury implies willingness to increase fall risk in order
to maintain an identity as normal
Calculated fall risk behavior
Adaptation to changing functional level
Risk taking beyond calculation
40
of the body could also be the cause of fear, especially when this was related to the fall causing the spinal cord injury:
so it’s turned into a fear of losing control, because it was the fact that I lost control of the moped that made me fall over the steering wheel….. That's the kind of things I’m left with, that deep fear. (woman, 32)
The majority of the informants took calculated risks but within safe limits, balancing benefits versus risks, in order to live their lives on their own terms. However, they were also willing to pay the price of falling in order to maintain a healthy self-image. Moving around without fall risk was looked upon as impossible because the alternative was some degree of inactivity and dependence.
I could of course have sat down in a wheelchair or, isolated myself in my own house or something like that, but that's because I try to be, live normally... Not that I seek out risk situations, I don’t feel I'm doing that... (man, 23)
Informants experiencing a deterioration of function after long lives with SCI, who were opposed to accepting the consequences of their poorer functioning, thus experienced and accepted an even higher fall risk. There were informants who were willing to take risks beyond the normal calculated ones. They exceeded both their own, peers‟ and family members‟ levels of comfort in order to achieve higher functional levels.
… I kind of have to challenge myself a little and… You don’t become a good skier without wiping out a few times. A lot of people would probably say that I’m a risk-taker… (man, 57)
ASSESSMENTS OF BALANCE CONTROL 4.5
BBS showed a ceiling effect as 28 per cent of the participants achieved the maximal sum score, whereas no floor or ceiling effect was observed for the Mini-BESTest (Figure 10). Both scales showed excellent internal consistency, with a Chronbach‟s alpha of ≥0.94. The correlation between Mini-BESTest and BBS was high (rs =0.90, p<0.001) confirming high criterion validity (Figure 10). The sum scores correlated highly with 10 MWT maximal speed, SCIM III mobility items and TUG (rs >0.70), showing high construct validity for both scales.
BBS and Min-BEStest were able to discriminate (known groups validity) community walkers not using walking aids (p<0.001) from individuals using mobility aids, with cut-off points at >47/56
41
Figure 10. Distribution of sum scores (A, B), Box and Whiskers plot with median and range of sum scores (C) and correlation (D) between sum scores on the Berg Balance Scale (sum score 0‒
56 points) and the Mini-BESTest (sum score 0‒28 points) (n=46). Each dot in Figure D represents one or more individuals.
on the BBS and >19/28 on the Mini-BESTest for community walkers without aids. The scales could also discriminate between participants with low (p<0.001) and high concerns about falling, with cut-off points of ≤46/56 on the BBS and ≤19/28 on the Mini-BESTest for high concern.
However, the specificity of the BBS in discriminating low and high concerns about falling was low (55%). None of the scales could discriminate between participants with infrequent and recurrent falls (p=0.78 for the BBS and p=0.64 for the Mini-BESTest) (Figure 11).
All possible response answers on the rating scales were used in both balance scales. However, there was a very low frequency (2–5%) for response categories 1 and 2 on the BBS (ranging from 0–4). Good person response validity was found for BBS and Mini-BESTest with two (4%) and
42
one (2%) participant, respectively, showing person misfit. The BBS was able to distinguish two groups with a separation index of 1.68 and a person reliability value of 0.74. The Mini-BESTest was able to distinguish more than three groups with a separation index of 2.95 and a person reliability value of 0.90. Person-item maps are shown in figures 12 and 13. For the BBS, one item was at the bottom (easiest item) of the logit scale, and no items were in the upper quarter (more difficult items) of the logit scale. The participants were spread over most of the logit scale; 46 per cent (n = 21) were in the upper quarter, and 7 per cent (n = 3) were in the lower quarter. For the Mini-BESTest items were spread from 25–75 per cent of the logit scale with no items at either the top or at the bottom, and the participants were spread across the entire logit scale.
4.5.1 Additional results
It is presumed that the BBS and Mini-BESTest both measure the construct of balance control. The correlation between the two scales was high in this study (rs =0.899, p<0.001). However, this correlation does not tell anything about the agreement between the two scales. To investigate the limits of agreement, the Bland‒Altman method was applied.103 The null hypothesis was that there would be no variation in the mean differences, which was tested with a one-sample t-test. This test showed a highly significant (p<0.001) mean difference, 19.3 per cent (CI; 22.9 to15.7), with BBS having the highest scores, thus indicating no agreement.
Figure 11. Receiver Operating Characteristic (ROC) curves of the Berg Balance Scale and the Mini-BESTest for classifying individuals according to A. walking ability (community walker not dependent on walking aids versus individuals using mobility aids) (n=46), B. low (FES I ≤ 27) versus high (FES-I >27) concerns about falling (n=46) and C. recurrent (>2) falls versus infrequent falls (≤2) (n=45).
False positive rate (100 – specificity)
A B C
Concerns about falling
True positive rate (sensitivity)
Walking ability Recurrent falls
44
Figure 12. Person-item map for the Bergs Balance Scale. Hierarchies of participant scores and item calibrations are displayed on a common logit scale which has been transformed to a possible range of 0 to 100. Each participant is presented by the symbol “x” on the left side of the scale (dotted vertical line). The distribution of items in order of difficulty is shown to the right of the scale. For both distributions median (M), one standard deviation (S) and two standard deviations (T) are indicated, for persons on the left side and for items on the right side of the scale.
Higher scoring persons
Lower scoring persons
More difficult
items
Less difficult
items
45
Figure 13. Person-item map for the Mini-BESTest. Hierarchies of participant scores and item calibrations are displayed on a common logit scale which has been transformed to a possible range of 0 to 100. Each participant is presented by the symbol “x” on the left side of the scale (dotted vertical line). The distribution of items in order of difficulty is shown to the right of the scale. For both distributions median (M), one standard deviation (S) and two standard deviations (T) are indicated, for persons on the left side and for items on the right side of the scale.
Higher scoring persons
Lower scoring persons
More difficult
items
Less difficult
items
46
47
5 DISCUSSION MAIN FINDINGS 5.1
Falls in individuals with SCI were frequent, and half of the total study sample reported recurrent falls in the previous year. Mode of mobility was an important factor associated with recurrent falls, as ambulatory individuals had higher odds than wheelchair users. Ability to get up from the ground, regular exercise and younger age were also associated with recurrent falls. A subgroup analysis indicated that the factors associated with recurrent falls differed for wheelchair users and ambulatory individuals.
Fifty-six (82%) of the ambulatory individuals reported falling during the 12-month follow-up, and 33 (49%) had at least three falls. Forty-four (65%) individuals experienced at least one injurious fall, with most injuries being minor. Three individuals, all menopausal women, had a serious injury. Reporting recurrent falls in the previous year and fear of falling were predictors of both recurrent falls and injurious falls in ambulatory individuals. In addition, slower walking speed was a predictor, although weak, for recurrent falls.
The ambulatory informants interviewed in the qualitative study considered falls to be a part of life. They were aware of their increased fall risk and used various preventive measures to avoid falls. Nevertheless, when important to their self-identity they were also willing to expose themselves to fall-risk situations. However, falls especially in every-day situations interfered with their identities and self-images as normal, healthy, and well-functioning individuals. All the informants expressed some conditional fall-related concerns i.e. walking on slippery surfaces, but a few had inexpedient concerns limiting their activities and participation in society.
Lack of balance was perceived as one of the main causes for falls. Yet, when measuring balance control with the BBS, no association with recurrent falls was found. Thus, a newer and presumably more sensitive balance scale, the BESTest, was tested. Both the BBS and Mini-BESTest proved to be valid scales for assessing balance control in individuals with chronic SCI who were able to walk. The Mini-BESTest seemed to be the preferable instrument in this sample due to its lack of ceiling effect and better scale properties. However, neither the BBS nor the Mini-BESTest could predict recurrent falls.